fix: handle resetter channel errors

[0x416e746f6e]

reset signal channels didn't check for errors on receive, and would only stop the subproxy if receive was Ok. however, it's possible that receiver lags, in which case .recv() would yield Err, which would cause the resetter task to exit w/o stopping the web-service.

consequently, this can cause one of the sub-proxies to remain active, while 2 others would stop, and the whole proxy would endlessly wait in the join_all.

this PR fixes above condition, improves error handling in general, and adds basic unit test.

[akundaz]

Have you tried using cancel tokens for this? It would completely eliminate the need to account for weird receiver behavior